home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.wizards
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!darwin.sura.net!wupost!newsfeed.rice.edu!exlogcorp!mcdowell
- From: mcdowell@exlogcorp.exlog.com (Steve McDowell)
- Subject: Re: Changing the owner of a process
- Message-ID: <1992Nov19.220759.1846@exlog.com>
- Keywords: process ownership
- Sender: mcdowell@exlog.com (Steve McDowell)
- Organization: EXLOG, Inc.
- References: <27551@dog.ee.lbl.gov><1992Oct29.162445.23551@eng.ufl.edu> <1cpjs4INNn3@early-bird.think.com> <1992Nov5.152833.27744@dale.ksc.nasa.gov> <1992Nov17.142837.21252@dale.ksc.nasa.gov>
- Date: Thu, 19 Nov 92 22:07:59 GMT
- Lines: 19
-
- In article <1992Nov17.142837.21252@dale.ksc.nasa.gov> torek@horse.ee.lbl.gov (Chris Torek) writes:
- > If you simply reach in and replace a process's uid, the kernel will
- > eventually panic when it notices the inconsistent counts.
-
- Why is it going to panic over an inconsistent process count? That's simply not good
- operating system design. When it hits an inconsistancy (as it eventually will
- in some remote region of the world) it should simply alert the operator and
- re-count things. If a process is facing starvation or scheduling temporarily becomes
- suspended because of the situation, then so what? It's still better than a panic.
-
- Panic's should be used strictly for irreversable situations; an out of whack counter
- shouldn't be an irreversable situation as long as the computer knows how to count. More
- attention needs to be paid to error recovery.
-
- --
- Steve McDowell . . . . o o o o o Opinions are
- Exlog, Inc. _____ o mine, not my
- mcdowell@exlog.com _____==== ]OO|_n_n__][. employers..
- [_________]_|__|________)<
-